Tunnel interface for securing traffic over a network

Copyright notice

Contained herein is material that is subject to copyright protection. The copyright owner has no objection to the facsimile reproduction of the patent disclosure by any person as it appears in the Patent and Trademark Office patent files or records, but otherwise reserves all rights to the copyright whatsoever. Copyright © 2000-2017, Fortinet, Inc.

Background

Field

Embodiments of the present invention generally relate to the field of Internet processors. In particular, embodiments of the present invention relate to methods and apparatus for delivering security services, such as firewalls.

Description of the Related Art

The service provider game has grown extremely crowded and fiercely competitive, with numerous players offering similar products and services. While having a large number of comparable services is arguably beneficial to the enterprise, it poses a host of potentially disastrous consequences for a service provider. If all competitors in a given market are offering services that are indistinguishable by the customer base, the burden of differentiation falls squarely on cost, with the least-cost competitor emerging “victorious”. Jockeying for the cost-leader position rapidly drives down service pricing, reducing margins to rubble and rendering the service a commodity. Furthermore, numerous offerings that are similar in attributes and cost make it very difficult to lock in customers.

Operational costs also present a significant challenge to service providers. Cumbersome, manual provisioning processes are the primary culprits. Customer orders must be manually entered and processed through numerous antiquated back-end systems that have been pieced together. Once the order has been processed, a truck roll is required for onsite installation and configuration of Customer Premises Equipment (CPE), as well as subsequent troubleshooting tasks. This is a slow and expensive process that cuts into margins and forces significant up-front charges to be imposed on the customer. In order to be successful in today's market, service providers must leverage the public network to offer high-value, differentiated services that maximize margins while controlling capital and operational costs. These services must be rapidly provisioned and centrally managed so that time-to-market and, more importantly, time-to-revenue are minimized. Traditional methods of data network service creation, deployment, and management present significant challenges to accomplishing these goals, calling for a new network service model to be implemented.

Basic Internet access, a staple of service provider offerings, has been commoditized to the point that margins are nearly non-existent. This fact has driven service providers to look for new value-added features and services to layer over basic connectivity so that they are able to differentiate on factors other than cost. The most significant opportunity for differentiation is found in managed network services. Managed network services enable enterprise IT organizations to outsource time-consuming tactical functions so that they can focus strategic core business initiatives.

Enterprise customers are now demanding cost-effective, outsourced connectivity and security services, such as Virtual Private Networks (VPNs) and managed firewall services. Enterprise networks are no longer segregated from the outside world; IT managers are facing mounting pressure to connect disparate business units, satellite sites, business partners, and suppliers to their corporate network, and then to the Internet. This raises a multitude of security concerns that are often beyond the core competencies of enterprise IT departments. To compound the problem, skilled IT talent is an extremely scarce resource. Service providers, with expert staff and world-class technology and facilities, are well positioned to deliver these services to enterprise customers.

While IT managers clearly see the value in utilizing managed network services, there are still barriers to adoption. Perhaps the most significant of these is the fear of losing control of the network to the service provider. In order to ease this fear, a successful managed network service offering must provide comprehensive visibility to the customer, enabling them to view configurations and performance statistics, as well as to request updates and changes. Providing IT managers with powerful Customer Network Management (CNM) tools bolsters confidence in the managed network service provider and can actually streamline the service provisioning and maintenance cycle.

Customer Premises Equipment (CPE)-Based Managed Firewall Services

Data network service providers have traditionally rolled out managed network service offerings by deploying specialized CPE devices at the customer site. This CPE is either a purpose-built network appliance that, in addition to providing specific service features, may also serve some routing function, or a mid to high-end enterprise-class server platform, typically UNIX-based. In the case of a managed firewall solution, the CPE device provides services that may include VPN tunnel termination, encryption, packet filtering, access control listings, and log files. The CPE at each customer site is aggregated at a multiplexer via leased lines and/or public Frame Relay PVCs (permanent virtual circuits) at the service provider POP (point of presence), then into a high-end access router and across the WAN (wide area network).

In many cases, service providers and enterprise customers find it too expensive and cumbersome to deploy CPE-based security at every site, but rather deploy secure Internet access points at one or two of the largest corporate sites. In this model, all remote site Internet traffic is backhauled across the WAN to the secure access point and then out onto the Internet, resulting in increased traffic on the corporate network and performance sacrifices.

Service providers face significant challenges when deploying, managing and maintaining CPE-based managed firewall services. When a customer expresses interest in utilizing such a service, a consultation with experienced security professionals is required to understand the corporate network infrastructure and site-specific security requirements, yielding a complex set of security policies. This may be accomplished through a set of conference calls or a number of on-site visits. Once the security requirements and policies have been identified, the service provider must procure the CPE device. In some cases, the equipment vendor may provide some level of pre-configuration based upon parameters supplied by the service provider. While CPE vendors are driving towards delivering fully templatized, pre-configured systems that are plug-and-play by enterprise staff, most service providers still assume the responsibility for on-site, hands-on configuration, and a truck-roll to each of the customer sites is necessary. This is particularly true in server-based CPE systems, where a relatively high degree of technical sophistication and expertise is required to install and configure a UNIX-based system.

Typically, a mid-level hardware and security specialist is sent onsite, along with an account manager, to complete the CPE installation and configuration. This specialist may be a service provider employee or a systems integrator/Value-Added Reseller (VAR) who has been contracted by the service provider to complete CPE rollout. This complex process begins with physical integration of the CPE device into the customer network. In the case of a CPE appliance, where the OS and firewall/VPN software components have been pre-loaded, the tech can immediately proceed to the system configuration phase. Server-based CPE services, however, require the additional time-consuming step of loading the system OS and software feature sets, adding a further degree of complexity.

In the configuration phase, the tech attempts to establish contact between the CPE device and central management system at the service provider NOC (network operations center). In cases where the device has not been previously assigned an IP address, an out-of-band signaling mechanism is required to complete the connection, typically a modem and a plain old telephone service (POTS) line. If the integration process has been successful, NOC staff should be able to take over the process, pushing specific policy configurations (and possibly an IP address) down to the CPE device through a browser-driven management interface. This entire process must be repeated for every enterprise site utilizing the managed-firewall service.

Additionally, maintenance processes and costs for CPE-based managed firewall services can also be overwhelming to both the service provider and enterprise customers. Enterprises are forced to either keep cold spares onsite or be faced with periods of absent security when their firewall fails, a situation that is unacceptable to most of today's information intensive corporations. Service providers must have an inventory of spares readily available, as well as staff resources that can, if necessary, go onsite to repeat the system configuration process. Troubleshooting thousands of CPE devices that have been deployed at customer sites is an extremely formidable challenge, requiring extensive call center support resources, as well technicians that can be quickly deployed onsite.

As CPE-based firewall services have traditionally been deployed in private enterprise networks, the original management systems for these devices have difficulty scaling up to manage several large, multi-site service provider customers. CPE device vendors are scrambling to ramp up these systems to carrier-grade and scale. Firewall management systems are typically GUI-based (graphical user interface-based), browser-driven interfaces that run on industrial grade UNIX platforms in the service provider NOC. The management system interfaces with the CPE devices based on IP address. The CPE-based managed firewall model faces service providers with another issue: capital costs. In addition to the significant costs required to build out a POP/access infrastructure, including multiplexers and high-capacity access routers, the service provider must also assume the initial costs of the CPE device, including firewall and VPN software licensing charges. In many cases, these costs are passed on to the customer. This creates steep up-front costs that, coupled with per-site installation charges, can present a serious barrier to service adoption. In markets where several service providers are offering managed firewall services, a service provider may absorb the CPE cost to obtain a price leadership position, cutting deeply into margins.

The CPE-based model is also limited when rolling out services beyond the managed firewall offering. New services, such as intrusion detection, may require additional hardware and/or software. This results in higher capital costs, as well as another expensive truck roll.

There is also a performance penalty in conventional IP-SEC-mode (Internet protocol secure mode) transmissions, in that each packet going through must be examined at the sending end of a transmission to determine whether it must be encrypted, and then each packet at the receiving end of the transmission to determine whether it must be decrypted.

Thus, there is a need for a method and apparatus of delivering a variety of network services, for example, security services, such as firewalls, and secure transmission of data across a network, such as the Internet.

Summary

Methods and systems are described for a flexible, scalable hardware and software platform that allows a service provider to easily provide Internet services, virtual private network services, firewall services and the like to multiple customers. According to one embodiment, a method is provided for delivering customized network services to subscribers of a managed security service provider. A request is received, at a service management system (SMS) of the managed security service provider, to establish an Internet Protocol (IP) connection between a first location of a first subscriber of multiple subscribers of the managed security service provider and a second location of the first subscriber. Responsive to the request, the SMS causes a tunnel to be established between a first service processing switch of the managed service provider that is associated with the first location and a second service processing switch of the managed service provider that is associated with the second location. The first service processing switch and the second service processing switch are coupled in communication via a public network.

Other features of embodiments of the present invention will be apparent from the accompanying drawings and from the detailed description that follows.

Brief description of the drawings

Embodiments of the present invention are illustrated by way of example, and not by way of limitation, in the figures of the accompanying drawings and in which like reference numerals refer to similar elements and in which:

FIG. 1 is a block diagram of a system having a plurality of ISP boxes connected to the Internet in accordance with an embodiment of the present invention.

FIG. 2 is a block diagram a service provider network in accordance with an embodiment of the present invention.

FIG. 3 is a block diagram of an IP service delivery platform in accordance with an embodiment of the present invention.

FIG. 4 is a block diagram of a system providing a plurality of virtual private networks in accordance with an embodiment of the present invention.

FIG. 5 is a block diagram of a ring-network hardware platform in accordance with an embodiment of the present invention.

FIG. 6 is a block diagram of a service processing switch in accordance with an embodiment of the present invention.

FIG. 7 is a block diagram of an integrated system including conventional existing network elements in accordance with an embodiment of the present invention.

FIG. 8 is a block diagram of hardware elements and software elements in accordance with an embodiment of the present invention.

FIG. 9 is a block diagram of a multiprocessor system using a ring network in accordance with an embodiment of the present invention.

FIG. 10 shows a block diagram of a system for comparison.

FIG. 11 shows a block diagram of a system for comparison.

FIG. 12 shows a block diagram of a system for comparison.

FIG. 13 shows a block diagram of a system for comparison.

FIG. 14 shows a block diagram of a system in accordance with an embodiment of the present invention.

FIG. 15 shows a block diagram a system in accordance with an alternative embodiment of the present invention.

Detailed description

Methods and systems are described for a flexible, scalable hardware and software platform that allows a service provider to easily provide Internet services, virtual private network services, firewall services and the like to a plurality of customers.

In the following detailed description of the preferred embodiments, reference is made to the accompanying drawings that form a part hereof, and in which are shown by way of illustration specific embodiments in which the invention may be practiced. It is understood that other embodiments may be utilized and structural changes may be made without departing from the scope of the present invention.

The leading digit(s) of reference numbers appearing in the Figures generally corresponds to the Figure number in which that component is first introduced, such that the same reference number is used throughout to refer to an identical component which appears in multiple Figures. Signals and connections may be referred to by the same reference number or label, and the actual meaning will be clear from its use in the context of the description.

In some embodiments, the present invention deploys one or more virtual private networks (VPNs) running on one or more carrier-class platforms that scale to provide cost-effective solutions for Internet service providers (ISPs). In particular, security services such as firewalls can be provided by the ISPs for services they provide to their customers, wherein a plurality of customers are hosted on a single network of processors. An ISP is providing hosting services (e.g., hosting an Internet web site for a customer) and routing (moving data to and from the Internet) for their customers.

FIG. 1 shows a system 100 that includes a plurality of similar ISP (Internet service provider) boxes 110 connected to the Internet 99 in accordance with an embodiment of the present invention. In this embodiment, each box 110 represents a subsystem having routing services provided by a block called access router 111 , and hosting services provided by blocks 113 and 114 . The ISP is typically a company that provides Internet services (such as connectivity to the Internet, as well as servers that store data and provide data according to requests by users, and network connectivity to users) to a plurality of customers including customer A and customer B. In some embodiments, customer premises equipment 117 and 118 (also called CPE 117 and 118 , this is hardware and the software that controls the hardware, that is installed at the customer's premises; this can include servers, routers and switches, and the network connecting to individual user's workstations, and various interfaces to external communications networks) is used to provide at least a portion of the function to support customers A and B respectively, and the ISP 110 provides the rest in blocks 113 and 114 respectively. The function to support customers includes such things as web site hosting, database and other servers, e-mail services, etc. The customer's CPE 117 and 118 connect to the ISP through, e.g., access router 111 and security services 112 to customer A site one 113 and customer B site one 114 , and also to the Internet 99 in a manner that isolates customer A and customer B from one another except for communications and E-mail that would normally pass across the Internet 99 .

Further, by establishing secure connections between two ISP boxes 110 across the Internet 99 , a virtual private network or VPN 410 (see FIG. 4 below) can be created. This function allows, for example, customer A's office at a first site (e.g., headquarters 117 ) to connect seamlessly to customer A's office at a second site (e.g., branch office 119 ) using what appears to them as a private network, but which actually includes some CPE at site 117 , some services 113 provided within ISP 110 . 1 , a secure encrypted connection across Internet 99 , some services also in ISP 110 . 2 , and some CPE at site 119 . Users at sites 117 and 119 can communicate with one another and share data and servers as if they were on a single private network provided by, e.g., VPN 410 .

FIG. 2 is a block diagram of a service provider (SP) network 200 in accordance with an embodiment of the present invention. A conventional network “cloud” 98 includes the SP's Internet protocol (IP) or asynchronous transfer mode (ATM) core, as is well known in the Internet art. IP system 201 connects to such existing infrastructure 98 , as well as to other optional conventional hardware such as described in FIG. 2 below, to provide SP network 200 . IP System 201 provides hardware 230 and software 220 to provide a plurality of virtual routers (VRs) 210 . Each VR 210 provides support for router services and server services such as those that provide customer site services 113 of FIG. 1 . Each VR 210 is supported by an object group 211 , which is a group of generally dissimilar objects such as routing object 212 , packet filtering object 213 , firewall object 212 , network address translation (NAT) object 215 , and/or other objects. In some embodiments, each VR 210 is a separate instantiation.

In some embodiments, software 220 includes IP network operating system (IPNOS) 223 , service management system (SMS) 221 (e.g., in some embodiments, this is the Invision™ software from CoSine Communications Inc., assignee of the present invention), and customer network management system (CNMS) 222 (e.g., in some embodiments, this is the Ingage™ software from CoSine Communications Inc., assignee of the present invention). SMS 221 provides such services as configuration of blades 239 , defining subscribers, determining services, and generation of IP security (IPSec) public/private key pairs. CNMS 222 provides such services as providing subscribers (customers) visibility to services. In some embodiments, CNMS software runs at least in part in a user's CPE or workstation, typically at a company's information services (IS) headquarters.

In some embodiments, IP server switch (IPSX) hardware 230 includes one or more scalable hardware enclosures, each having a plurality of service “blades” 239 (i.e., an insertable and removable printed circuit card having one or more processors, each having its own CPU and memory) each connected in a ring configuration (such as a counter-rotating dual ring 232 ). In some embodiments, three types of blades 239 are provided: control blade(s) 234 , processor blade(s) 236 , and access blade(s) 238 . IPSX hardware also includes highly available, redundant, and hot-swappable hardware support 240 including power supplies 241 and fans 242 .

FIG. 3 is a block diagram of an IP service delivery platform (IPSDP) 300 in accordance with an embodiment of the present invention. The hardware and software of SP network 200 can be viewed as generating various network “clouds” such as edge cloud 95 , access concentration cloud 96 , and service processing cloud 97 . These are built upon the existing conventional SP's IP or ATM core cloud 98 and they connect to the external Internet cloud 99 . IPSDP 300 includes an ISP's SP network 200 connected to one or more customer's offices 301 each of which includes some amount of CPE 110 . In the embodiment shown, three corporate remote offices 301 . 1 are connected to SP network 200 using various conventional communications devices, well known to the art, such as frame relay switch 326 , M13 multiplexor (mux) 327 , DSLAM (digital subscriber link access multiplexor) 328 , and dial-up RAS (remote access server) 329 (used to receive dial-up connections, for example, from the modem 311 connected to laptop computer 316 in portable system 310 of dial-up telecommuter 301 . 3 ). In the embodiment shown, SP network 200 includes two systems 201 , on connecting through frame relay switch 326 , M13 multiplexor (mux) 327 , DSLAM 328 , and dial-up RAS 329 to remote office's CPE 110 , and the other connecting directly to the customer's corporate headquarter's CPE 1110 (which also includes a control and monitoring function provided by CNMS 222 ) using conventional communications protocols such as frame relay (FR, an access standard defined by the ITU-T in the 1.122 recommendation “Framework for Providing Additional Packet Mode Bearer Services”), Internet protocol (IP), FT1 (fractional T1), T1/E1 (a digital transmission link with capacity of 1.544 Megabits per second), FT3 (fractional T3), T3 (capacity of 28 T1 lines), and/or OC3 (optical carrier level 3=three times the OC1 rate of 51.840 Mbps)(each of which is a conventional communications service well known to the art).

In some embodiments, IPDSP 300 provides a VPN 410 , using secure connections across the Internet 99 , to connect remote offices 301 to one another.

FIG. 4 is a block diagram of a system 400 providing a plurality of virtual private networks 410 , 420 , 430 , 440 in accordance with an embodiment of the present invention. VPNs 420 , 430 , and 440 are each equivalent to the VPN 410 that supports subscriber 1 , except that they are for other subscribers. Each subscriber has a set of partitioned virtual routers 210 . For example, subscriber 1 has two locations, 411 and 412 , connected in a VPN 410 . VR 210 at location 411 can include some CPE 110 as well as support provided in system 201 - 1 . VR 210 at location 412 can include some CPE 110 as well as support provided in system 201 - 2 . These two VRs 210 establish a “tunnel,” a secure connection, that allows them to maintain secure communications that support the VPN 410 even across packet networks such as the Internet 99 . Each VR 210 is the equivalent of an independent hardware router. Since each VR 410 is supported by an object group 211 , objects can be easily added or omitted to enable customized services on a subscriber-by-subscriber basis to meet each subscriber's individual needs. SMS 221 running on SP network 200 allows ease of service provisioning (dynamically adding additional processors/processing power when needed, reducing the processors/processing power used for VPN 410 when not needed). In some embodiments, IPNOS 223 uses an open Application Program Interface (API) to enable new services to be added to the platform whenever needed.

In some embodiments, system 401 at a first site (e.g., an ISP premises locally connected to a customer office) includes IPSX 201 - 1 having a VR 210 connected to CPE 117 . This system 401 appears to the outside world as a single router having firewall services, server(s) and user(s), etc. These functions can be provided by either or both VR 210 and CPE 117 , thus allowing a customer to outsource many or most of these services to the service provider and IPSX 201 - 1 . Similarly, system 402 at a second site (e.g., another ISP premises locally connected to a remote office of the same customer) includes IPSX 201 - 2 having a VR 210 connected to CPE 119 . This system 402 also appears to the outside world as a single router having firewall services, server(s) and user(s), etc. These functions can be provided by either or both VR 210 and CPE 119 , thus allowing a customer to outsource many or most of these services to the service provider and IPSX 201 - 2 .

FIG. 5 is a block diagram of a ring-network hardware platform 230 in accordance with an embodiment of the present invention. Hardware platform 230 includes plurality of service “blades” 239 (i.e., an insertable and removable printed circuit card having one or more processors, each having its own CPU and memory) each connected in a ring configuration (such as a counter-rotating dual ring 232 ). In some embodiments, three types of blades 239 are provided: control blade 234 (not shown here), processor blades 236 (providing such functions as point-to-point (PPTP) connectivity, firewall protection against hacking, intruders, or accidental access), and access blades 238 (providing such functions as NAT, encryption, and routing).

FIG. 6 is a block diagram of a service processing switch 600 in accordance with an embodiment of the present invention. In some embodiments, service processing switch 600 includes a hardware enclosure 230 having power supplies 241 that are hot-swappable, redundant, capable of automatic failover (when one fails, others take over), and which can be AC or DC sourced. In some embodiments, dual hot-swappable, variable speed fans 242 are provided. In some embodiments, software updates can be made without system downtime by swapping out all object groups 211 (virtual routers), changing the software modules, and then resuming processing. In some embodiments, all service blades 239 are hot-swappable (they can be removed and/or inserted without bringing the system down) and include automatic failover from primary mode to protect mode. In some embodiments, dual counter-rotating rings 232 support primary and protect redundancy. In some embodiments, system 600 provides NEBS Level 3 compliance and is Y2K ready, provides SONET (synchronous optical network) 1+1 Line Protection Switching, and includes integrated metallic cross-connects to enable DS3 (digital signal level 3; 44,736,000 bits per second) blade automatic failover without touching the facility.

FIG. 7 is a block diagram of an integrated system 700 including conventional existing network elements in accordance with an embodiment of the present invention. Integrated system 700 optionally includes conventional frame relay switch 326 , M13 mux 327 , Digital Subscriber link Access Multiplexor (DSLAM) 328 , and Remote Access Server (RAS) 329 connecting to customer's equipment such as CPE router 110 and dial-up system 310 . In some embodiments, integrated system 700 optionally includes a core IP router 720 and/or a core ATM switch as part of an SP core 98 . This provides support for a large number of conventional technology standards, and interoperability with existing access-concentration and core-network elements. It also offers interworking between frame-relay networks and IP networks. Network address translation (NAT) enables enterprise subscribers to leave their network addressing untouched. It also enables one to merge IP and legacy networks into one, with continuity of service (COS) guarantees.

FIG. 8 is a block diagram of hardware elements 230 and software elements 220 in accordance with an embodiment of the present invention. Hardware elements 230 include a 26-slot, two-sided chassis 831 having a 22-gigabit per second (Gbps) ring midplane 832 . Service blades 239 can be hot plugged into midplane 832 form either side of chassis 831 . Three types of service blades 239 are provided: control blades 234 , processor blades 236 , and access blades 238 . In some embodiments, four processors are provided on each service blade 239 , each processor having a CPU and its own memory, allowing specialized processing to be performed on various different daughter cards of the blades 239 .

In some embodiments, a single system chassis 831 provides a redundant back plane and blade-termination facilities 832 . The access blades 238 , processor blades 236 , control blades 234 , power supplies 241 and fan trays 242 are designed for hot-swappable operation—any of these components may be removed from service while the entire system remains operational. The metallic cross connect is a passive system that provides fail-over support to allow DS3 and DS1 access facilities to be switched from one access blade to another access blade should an access port or card fail. The phase 1 chassis provides 26 universal slots, each of which may be populated with control blades, access blades, and processor blades. To operate, the chassis must contain at least one control blade. Up to two control blades may be operational in a chassis at the same time. Access blades are added as input/output requirements grow, and processor blades are added as computation requirements scale.

In some embodiments, each system 230 supports up to twenty-five processing blades (PB) 236 . Each processor blade 236 is designed to support three hundred Mbps of full duplex traffic while delivering IP services including application firewall, LT2P, PPTP, NAT, VPN router.

In some embodiments, each system 230 supports up to two control blades (CB) 234 . CBs 234 provide overall system supervision, IP route calculation, software update management, and network management statistics logging services. When two CBs 234 are operational within a chassis 831 , they remain synchronized such that should either CB 234 fail, the other CB 234 automatically takes over system operation. In this process all active services remain in progress. Each control blade 234 is hot swappable, so that when proper procedures are followed, a failed or malfunctioning CB 234 may be removed from on operational system 230 without bringing down any customer services.

In some embodiments, each CB 234 provides four Ethernet interfaces for management traffic. Each Ethernet interface has a distinct collision domain and may each be configured with a primary and secondary IP address. Ethernet interfaces designated for management use may be configured for primary and protected configurations, both sharing the same IP address, reducing ISP IP address requirements. The CB 234 Ethernet interfaces may be configured for fully meshed communications over diverse paths to diverse operating systems. Each CB 234 is also equipped with a random # seed generator for use in security applications.

In some embodiments, each system 230 supports up to twenty-five access blades (AB) 238 . Access blades 238 provide physical line termination, hardware-assisted IP forwarding, hardware assisted encryption services, and hardware assisted queue management. Each access blade 238 is hot swappable, so that when proper procedures are followed, a failed or malfunctioning access blade may be removed from on operational system 230 without bringing down any customer services. In some embodiments, 10/100 Ethernet-, DS3-, and OC3-type access blades are supported by system 230 .

FIG. 9 is a block diagram of a multiprocessor system 900 using ring network 932 in accordance with an embodiment of the present invention. In some embodiments, each of two network rings 933 and 934 connect nodes 931 together, where each blade 239 includes one or more nodes 931 , and each node 931 is connected to one or more processors 930 . In some embodiments, each processor is a high-performance processor such as an R12K processor from MIPS Corporation. In one embodiment, each blade 239 includes four nodes 931 , each having one processor 930 . Each processor 930 includes its own CPU (central processing unit) 935 and memory 936 , and optionally includes other hardware such as routers, encryption hardware, etc. Software tasks, in some embodiments, are split up such that one processor operates on one part of the data (e.g., the Level 7 processing) and another processor operates on another part of the data (e.g., the Level 3 processing). In other embodiments, the various processing portions of a task all run on a single processor, multiprocessing with other tasks that share that processor. Thus, the hardware provides scalability, where low-end systems include few processors that do all the work, and high-end systems include one hundred or more processors and the work is distributed among the processors for greater speed and throughput. In some embodiments, the plurality of processors 930 in the ring configuration includes forming dual counter rotating ring connections 933 and 934 , each connecting to each of the plurality of processors 930 .

In some embodiments, a separate control ring 935 is provided, connected to all processors 930 . Data passed on the control ring 935 allows control communications to be passed between processors, and in particular, allows the control blade to configure and control the other blades in IPSX 201 . In other embodiments, control ring 935 is omitted, and its function is overlaid on rings 933 and 934 .

Logical Queue Identifiers

In some embodiments, rings 933 and 934 are packet-passing rings. Each packet 950 placed in the rings includes a data portion 953 and a processor element identifier (PEID 951 ) that identifies for each node 931 which processor that packet is destined for, for example a 16-bit PEID that specifies one of 65526 PEs. If the PEID matches a processor on its particular node, the node 931 passes the packet to the proper processor 930 ; if not, the packet is forwarded to the next node 931 . In some embodiments, each packet also includes a logical queue identifier (LQID) that identifies a software entity (for example, an object group for a particular VR 210 ) residing on that processor 930 for which the packet is destined.

In some embodiments, every node 931 has a unique, globally unique (i.e., unique within an IPSX 201 , or within an ISP having a plurality of IPSXs 201 ) PEID 951 . In some embodiments, the way this is done is that one takes the blade ID (e.g., five bits) and you append the PE number, which is, for example, a eleven bits. Put that together in some fashion and you'll get a unique ID that is globally unique within some hardware configuration. Note that packets including this PEID 951 are routable. Just by looking at the PEID 951 , the system 201 has a topological structure so that it can route based on purely the PEID 951 . The next thing to keep in mind is that system 201 is managing multiple virtual context. Each VR 210 in a system 201 is a virtual router to which packet are to be directed. When packets come into node N 931 for example, system 201 needs to be able to steer it to the appropriate logical entity, i.e., to the appropriate context and to the object channel that it represents. Thus, a logical queue ID 952 is appended that is unique within the destination processor (PE) 930 . If an object in a processor 930 on node 1 930 wants to set up a channel to another object a processor 930 on node N 930 , they need to use the LQID 952 . A first LQID 952 and PEID 951 together represent the local end, and a second LQID 952 and PEID 951 together represent the remote end of the object and so the system can map the corresponding object channel, defining the object channel that is going across the network. From a networking perspective, PEID 951 looks like your IP address that routes packets like an IP address. But once you go to a particular node 931 , the LQID looks like the UDP (User Datagram Protocol, a TCP/IP protocol describing how messages reach programs within a destination computer) code number. So system 201 (e.g., SMS 221 ) essentially signals and negotiates the proper LQID to have a channel going between those ends. This allows all the traffic coming into a PE 930 to be steered along the appropriate object path to the appropriate object channel on that object.

In some embodiments, an object could be talking on another channel to another object, or to the same object, using a different channel. In which case each channel uses a different LQID 952 , but the same PEID 951 .

In some embodiments, system 201 sets up a shortcut that circumvents traffic that otherwise would be transmitted outside system 201 and then back in (e.g., traffic between two different VRs 210 supporting different customers). To set up such a shortcut, system 201 allocates a different LQID 952 for the shortcut. Thus, an object channel has the normal point-to-point path for normal traffic and has a multi-point-to-point path, which is used for shortcut traffic. So when packets come in to the object it knows whether the packet came in on the normal path or on the shortcut path. Similarly, when the object wants to use a shortcut, it also needs to allocate a different LQID for its outbound shortcut traffic. One interesting distinction of shortcut paths is that the normal point-to-point is bidirectional and data can flow in both directions, but shortcuts data flow flows in only one direction. So a receive site can have any number of transferred sites. Any number of objects can be transmitting to the same receive site. That is why it is called multi-point-to-point.

Further, some embodiments have different levels of shortcuts. For example, a packet can be sequentially passed to successive destinations in some embodiments. Thus there can be a complex multistage path. The shortcuts can trickle down to the ultimate end, where the packet cascades. Further, if one object knows a shortcut, it can tell other objects about its shortcut. So the other object does not have to come to the first object and then be directed to the shortcut destination, but rather can directly use the shortcut it has learned about.

The description continues in the full USPTO document.